A low-power, high-speed and high-density negative differential resistance (NDR) based (NDR-based) SRAM cell which can provide DRAM-like densities at SRAM-like speeds has been proposed by Farid Nemati and James D. Plummer in “A Novel High Density, Low Voltage SRAM Cell with a Vertical NDR Device,” 1998 Symposium on VLSI Technology Digest of Technical Papers, IEEE, pages 66-67, 1998.
The memory device structure is shown by FIG. 1 and is designated by reference numeral 10; the memory device structure is called a Thyristor-based Random Access Memory (T-RAM) cell. The T-RAM device or memory cell 10 consists of a thin vertical pnpn thyristor 12 with a surrounding nMOS gate 14 as the bistable element and a planar nMOSFET as the access transistor 16. The circuit schematic of the T-RAM cell 10 is shown by FIG. 2.
To access the T-RAM cell 10, two wordlines are necessary. The first wordline WL1 is used to control an access gate of the transfer nMOSFET device 16, while the second wordline WL2 is the surrounding nMOS gate 14 which is used to control the switch of the vertical pnpn thyristor 12. The thyristor 12 is connected to a reference voltage Vref. The second wordline WL2 improves the switching speed of the thyristor 12 from 40 ns to 4 ns with a switching voltage. A bitline BL connects the T-RAM cell 10 to a sense amplifier for reading and writing data from and to the T-RAM cell 10. The T-RAM cell 10 exhibits a very low standby current in the range of 10 pA.
When writing a “high”, the bitline BL is set at low, and both wordlines WL1, WL2 are switched on. At this moment, the thyristor 12 behaves like a forward biased pn diode. After a write operation, both gates are shut off, and a “high” state is stored in the thyristor 12. In a read operation, only the first wordline WL1 is activated, a large “on” current will read on the bitline BL through the access gate. When writing a “low”, the bitline BL is set at “high” state, and both wordlines WL1, WL2 are switched on. At this moment, the thyristor 12 behaves like a reverse biased diode. After the write operation, both gates are shut off, and a “low” state is stored in the thyristor 12. Similarly, in a consequence read, a very low current will be detected on the bitline BL. Further details of the operation of the T-RAM cell 10 and its gate-assisted switching are described in Nemati et al.; the contents of which are incorporated herein by reference.
A T-RAM array having a plurality of T-RAM cells 10 has demonstrated a density equivalent to that of DRAM arrays and a speed equivalent to that of SRAM arrays. Hence, the T-RAM array provides advantages afforded by both SRAM and DRAM arrays. These advantages make T-RAM an attractive choice for future generations of high speed, low-voltage, and high-density memories and ASICs.
However, there are several drawbacks of the T-RAM cell 10. First, there is the requirement of forming the thyristor 12 having a vertical pillar on a substrate during a fabrication process. Difficulties arise in controlling the dimensions of the vertical pillar and reproducing these dimensions for each T-RAM cell 10 in the T-RAM array. Second, due to the existence of a vertical thyristor 12 in each T-RAM cell 10, each T-RAM cell 10 is not planar and therefore difficult to scale. Third, it is difficult to control the dimension while forming the surrounding base gate around the base of each vertical thyristor 12. Fourth, each T-RAM cell is fabricated prior to or after fabricating any other devices, such as p-MOS and n-MOS support devices (i.e., sense amplifiers, wordline drivers, column and row decoders, etc.), which results in extra fabrication steps, thereby increasing thermal budget and manufacturing cost. Finally, due to these drawbacks, the resulting T-RAM cell 10 cannot be smaller than 8F2 and the cost of fabricating a T-RAM array is high.